Main » News and comments » » Artificial Intelligence in 2024

Artificial Intelligence in 2024

12.02.2024
672

https://nationalinterest.org/blog/techland/next-chapter-artificial-intelligence-209250

Artificial intelligence (AI) has emerged as a major driver of stock market gains in 2023, despite concern and anxiety over the technology's unchecked status. The fear that it will turn us into prey for Terminator-type robots has faded, although not completely. What worries modern experts is China's ambition to become a leading country in the field of AI, as it published more than twice as many scientific papers in 2022 in AI than the United States, while increasing its advantage over the United States in AI-related applications. However, all indications are that the United States will remain the dominant global power in artificial intelligence and machine learning in 2024.

According to Tracxn Technologies, there are now more than 18,000 artificial intelligence startups in the United States. A recent Adobe study found that 77 percent of consumers are now using artificial intelligence technologies. Moreover, consulting firm McKinsey found that AI adoption in business has more than doubled since 2017, with half of all U.S. businesses using some form of AI. Overall, McKinsey projects that generative AI alone will add between $2.6 trillion and $4.4 trillion a year to the global economy, with much of that coming from the United States, the world's dominant AI power.

America's dominance in artificial intelligence did not happen by accident, nor was it the result of government planning and investment, such as China's attempts to catch up with the United States under President Xi Jinping. A well-known artificial intelligence optimist, Lisa Su, CEO of American Micro Devices (AMD), believes that the overall business climate is key to US dominance in AI. AMD was one of the pioneers of semiconductor innovation, giving Silicon Valley its name. However, although AMD had a relatively stable market for gaming chips in the 2000s, its market dominance was eclipsed by companies such as TSMC, Intel, Qualcomm and Nvidia. In 2014, the company was on the verge of bankruptcy. Then Lisa came. Her leadership led to the introduction of a new generation of cutting-edge processors for AMD customers, called the Ryzen generation, culminating with the release of the AMD MI300X chip in November of this year. Su said the new chip is "the most advanced artificial intelligence accelerator in the industry."

It was a chip aimed directly at the industry leader in artificial intelligence chips, Nvidia. Su's entire presentation emphasized the MI300X's superiority over Nvidia's H100 chip, which companies like Amazon, Meta, Microsoft and Google are using to develop their own generative AI tools.

The move to AI chips is a good move for AMD, as gaming chips are exactly the type of chips needed for AI-focused chips. Unlike standard computer processors (CPUs), graphics processing units (GPUs) are equipped with thousands of cores that accelerate the machine learning process that powers generative artificial intelligence.

GPUs are used to develop and improve artificial intelligence algorithms. Another set of microchips, field programmable gate arrays (FPGAs), apply trained artificial intelligence algorithms to real data input to solve specific problems. Once manufactured, FPGAs can be field programmed for specialized computing applications.

The AMD MI300X chip contains both blocks, including application-specific integrated circuits (ASICs) specializing in artificial intelligence. Su expects the new chip to reach $1 billion in sales by mid-2024 (the MI300A variant will serve the supercomputer market).

Gaining a foothold in this market means winning over major buyers such as Microsoft and Meta. Both companies were present at the chip's unveiling, as well as OpenAI, which will use AMD's new chips in the latest version of its Triton AI software. Whatever happens to AI as a technology, thanks to companies like AMD, the competition to provide the means to do it faster, cheaper and more efficiently will be fiercer than ever.

 

Read also:

Large Hadron Collider Researchers Propose Building a Larger Superaccelerator

All 27 EU Countries Approved 50 billion Euros Aid to Ukraine